A ground-to-air integrated operation status monitoring and early warning system
By adjusting the fracture threshold, based on the difference and change evaluation value of traffic data, the accuracy of traffic operating conditions warning is improved, and the problem of poor clustering effect caused by inappropriate settings of fracture thresholds in the prior art is solved.
Patent Information
- Application Number
- CN202410549123.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-06
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-05-06
AI Technical Summary
The existing dynamic split clustering algorithm for connecting graphs is not suitable when setting the fracture threshold, resulting in poor clustering of traffic data, thereby reducing the accuracy of early warning of traffic operation conditions.
By obtaining traffic data and location information of different nodes on the road, a connection diagram is constructed, traffic characterization values are calculated, time periods are divided, vehicle change evaluation value is evaluated, and fracture thresholds are adjusted according to differences to improve the accuracy of cluster clusters.
It improves the accuracy of traffic operation conditions warning, ensures the accuracy of clustering effects, and reduces false alarms caused by factors such as traffic lights.
Smart Images

Figure CN118351694B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of traffic operation status early warning, and in particular to a ground-to-air integrated operation status monitoring and early warning system. Background Art
[0002] Smart city is a new form of city built by low-carbon urban ecosystem with more scientific development, more efficient management and better life. As an important part of smart city, the construction of smart transportation greatly affects the travel experience of urban residents. When building smart transportation, it is necessary to combine the interaction between smart networked vehicles and roads. Generally, traffic information on the road is obtained first, and the vehicle information is summarized through the nodes obtained from the traffic information on the road. The vehicle information between the nodes is analyzed to warn the operation status of the traffic, avoid long-term traffic jams that affect the travel experience of residents, and provide good conditions for smart transportation.
[0003] The existing dynamic splitting clustering algorithm of the connected graph is generally used to analyze the traffic data on the road, determine the abnormal situation of the traffic data, and issue an early warning of the traffic operation status according to the abnormal situation. However, when the dynamic splitting clustering algorithm of the connected graph is used to construct the connected graph, the break threshold will affect the clustering effect. If the break threshold is not set appropriately, the clustering effect of the traffic data will be poor, and the early warning accuracy of the traffic operation status will be low. Summary of the invention
[0004] In order to solve the problem that the existing dynamic splitting clustering algorithm of the connected graph has an inappropriate setting of the fracture threshold when clustering the traffic data on the road, resulting in low warning accuracy of the traffic operation status, the purpose of the present invention is to provide a ground-air integrated operation status monitoring and early warning system, and the technical scheme adopted is as follows:
[0005] The present invention provides a ground-to-air integrated operation status monitoring and early warning system, comprising a memory and a processor, wherein the processor executes a computer program stored in the memory to implement the following steps:
[0006] Obtaining traffic data of different nodes on the road at each moment in each cycle in the current time period and location information between different nodes; constructing a connectivity graph based on the traffic data and location information between different nodes;
[0007] According to the difference between the traffic data of the vehicles passing through each node at each moment and the traffic data of the vehicles passing through each node at each moment, the traffic characterization value of each node at each moment is obtained; based on the traffic characterization value of the moment in the time neighborhood of each node at each moment, each cycle is divided to obtain the sub-time period corresponding to each node in each cycle;
[0008] Based on the traffic characterization value of each node at all times in each cycle, the data sequence corresponding to each node in each cycle is constructed, and the data sequences corresponding to different nodes in the same cycle are matched. According to the matching results, the difference in the length of the sub-time period at each moment and other sub-time periods in the cycle, and the discrete degree of the traffic characterization value of all moments in the time neighborhood of each node at each moment, the vehicle change evaluation value of each node at each moment is obtained;
[0009] According to the difference in traffic characterization values of each node and the nodes connected to it at the same time and the vehicle change evaluation value, the fracture threshold corresponding to each node is adjusted to obtain the adjusted fracture threshold corresponding to each node at each time, the connected graph is divided into clusters based on the adjusted fracture threshold, and whether to issue an early warning is determined based on the cluster division result.
[0010] Preferably, obtaining the traffic characterization value of each node at each moment according to the difference between the traffic data of the vehicles passing through each node at each moment and the traffic data of the vehicles passing through each node at each moment in its time neighborhood includes:
[0011] For the jth node at the i-th moment:
[0012] The vehicle passing through the jth node at the i-th moment is recorded as the target vehicle at the j-th node at the i-th moment;
[0013] The traffic characterization value of the jth node at the i-th moment is calculated using the following formula:
[0014]
[0015] Among them, A ij represents the traffic characterization value of the jth node at the i-th moment, N ij represents the number of target vehicles at the jth node at the i-th moment, N ij represents the number of target vehicles at the jth node at the i-th moment, a ijn represents the traffic data of the nth target vehicle at the i-th time of the j-th node, M represents the number of moments in the time neighborhood of the i-th time, and a' ijn,m represents the traffic data of the nth target vehicle at the jth node at the i-th moment in the time neighborhood of the i-th moment, and exp() represents an exponential function with a natural constant as the base.
[0016] Preferably, the step of dividing each cycle to obtain a sub-time period corresponding to each node in each cycle based on the traffic characterization value of each node in the time neighborhood at each time, respectively, includes:
[0017] For the i-th moment:
[0018] The moment before the i-th moment in the time neighborhood of the i-th moment is recorded as the first neighboring moment of the i-th moment, and the moment after the i-th moment in the time neighborhood of the i-th moment is recorded as the second neighboring moment of the i-th moment;
[0019] For the jth node: the sum of the traffic characterization values of all the first neighboring moments of the jth node at the i-th moment is recorded as the first eigenvalue of the jth node at the i-th moment; the sum of the traffic characterization values of all the second neighboring moments of the jth node at the i-th moment is recorded as the second eigenvalue of the jth node at the i-th moment; the variance of the traffic characterization values of all the first neighboring moments of the jth node at the i-th moment is recorded as the first variance of the jth node at the i-th moment; the variance of the traffic characterization values of all the second neighboring moments of the jth node at the i-th moment is recorded as the second variance of the jth node at the i-th moment; according to the first eigenvalue, the second eigenvalue, the first variance and the second variance, the segmentation index of the jth node at the i-th moment is obtained, the first eigenvalue and the second eigenvalue are both positively correlated with the segmentation index, and the first variance and the second variance are both negatively correlated with the segmentation index;
[0020] Based on the segmentation index of each node at each moment, the moments in the cycle within the current time period are divided to obtain the sub-time periods corresponding to each node in each cycle.
[0021] Preferably, the step of dividing the moments in the cycle within the current time period based on the segmentation index of each node at each moment to obtain the sub-time periods corresponding to each node in each cycle includes:
[0022] For the j-th node: the moment when the segmentation index is greater than the preset segmentation threshold is taken as the division point, and the moments in each cycle in the current time period are divided based on the division point to obtain the sub-time periods corresponding to the j-th node in each cycle.
[0023] Preferably, a dynamic time warping algorithm is used to match data sequences corresponding to different nodes in the same period.
[0024] Preferably, the vehicle change evaluation value of each node at each moment is obtained according to the matching result, the difference in duration between the sub-time period at each moment and other sub-time periods in its cycle, and the discrete degree of the traffic characterization values of all moments in the time neighborhood of each node at each moment, including:
[0025] For the i-th moment:
[0026] In the cycle where the i-th moment is located, with a step length of 2, take the i-th moment as the starting point, traverse before and after the i-th moment respectively, obtain the corresponding moments, and use all the corresponding moments obtained as the reference moments of the i-th moment;
[0027] For the jth node: if the number of matching points of the traffic characterization value of the jth node at the i-th moment in the matching results of the data sequence corresponding to the jth node in the cycle at the i-th moment and the data sequence corresponding to the rth node in the cycle at the i-th moment is 1, then the matching point is used as the matching point corresponding to the jth node in the r-th node; if the number of matching points of the traffic characterization value of the jth node at the i-th moment in the matching results of the data sequence corresponding to the jth node in the cycle at the i-th moment and the data sequence corresponding to the r-th node in the cycle at the i-th moment is greater than 1, then the matching point with the shortest time interval with the i-th moment among all the matching points is used as the matching point corresponding to the jth node in the r-th node;
[0028] According to the difference in duration between the sub-time period at the i-th moment and the sub-time period at the q-th reference moment of the i-th moment, the traffic characterization value of the j-th node at the i-th moment, the traffic characterization value of the matching point corresponding to the j-th node in the r-th node, and the variance of the traffic characterization values of all moments in the neighborhood of the j-th node at the i-th moment, the vehicle change evaluation value of the j-th node at the i-th moment is obtained.
[0029] Preferably, the vehicle change evaluation value of the jth node at the i-th moment is calculated using the following formula:
[0030]
[0031] Among them, W ij represents the vehicle change evaluation value of the jth node at the i-th moment, Q represents the number of reference moments at the i-th moment, and l i Indicates the duration of the sub-time period at the i-th moment, l i,q represents the duration of the sub-time period where the qth reference moment of the i-th moment is located, R represents the number of nodes, and f ij represents the traffic characterization value of the j-th node at the i-th moment, represents the traffic characterization value of the matching point corresponding to the jth node in the rth node, ε i,j represents the variance of the traffic characterization values at all times in the neighborhood of the j-th node at the i-th time, || represents the absolute value sign, and exp() represents an exponential function with a natural constant as the base.
[0032] Preferably, the adjusting the fracture threshold corresponding to each node according to the difference in traffic characterization values of each node and the nodes connected to it at the same time and the vehicle change evaluation value to obtain the adjusted fracture threshold corresponding to each node at each time includes:
[0033] Based on the vehicle change evaluation values of each node at all times in each cycle, a vehicle change evaluation value sequence corresponding to each node in each cycle is constructed; a dynamic time warping algorithm is used to match the vehicle change evaluation value sequences corresponding to different nodes in the same cycle to obtain matching data of the vehicle change evaluation value of each node at each time;
[0034] For the jth node at the i-th moment:
[0035] The node connected to the jth node is recorded as the reference node of the jth node;
[0036] For the h-th reference node: the difference between the traffic characterization value of the j-th node at the i-th moment and the traffic characterization value of the matching point corresponding to the j-th node in the h-th node is recorded as the first difference corresponding to the h-th reference node of the j-th node;
[0037] The sum of the first differences corresponding to all reference nodes of the j-th node is recorded as the third eigenvalue of the j-th node; the variance of the matching data of the vehicle change evaluation values of all reference nodes of the j-th node at the i-th moment is calculated, recorded as the third difference;
[0038] Obtaining an adjustment parameter of the j-th node at the i-th moment according to the third eigenvalue and the third difference, wherein the third difference is positively correlated with the adjustment parameter, and the third eigenvalue is negatively correlated with the adjustment parameter;
[0039] An adjusted fracture threshold corresponding to the j-th node at the i-th moment is determined based on the adjustment parameter.
[0040] Preferably, the determining the adjusted fracture threshold corresponding to the j-th node at the i-th moment based on the adjustment parameter includes:
[0041] The sum of the adjustment parameter of the jth node at the i-th moment and the constant 1 is recorded as the adjustment coefficient of the jth node at the i-th moment, and the product of the adjustment coefficient and the initial fracture threshold is taken as the adjusted fracture threshold corresponding to the jth node at the i-th moment.
[0042] Preferably, judging whether to issue an early warning based on the clustering result includes:
[0043] For any cluster: if the number of data points in the cluster is less than a preset first number, the cluster is regarded as an abnormal cluster;
[0044] Calculate the sum of the number of data points in all abnormal clusters and the number of data points not in the clusters; if the sum is greater than a preset second number, issue a warning; if the sum is less than or equal to the preset second number, do not issue a warning; the preset first number is less than the preset second number.
[0045] The present invention has at least the following beneficial effects:
[0046] The present invention takes into account that when the connected graph dynamic split clustering algorithm is used to cluster the traffic data on the road, the inappropriate setting of the fracture threshold will directly affect the clustering effect, thereby resulting in low accuracy of early warning of traffic operation conditions, and high similarity of traffic conditions on the same road section in the same time period every day. Therefore, the present invention is based on the traffic characterization value of each moment in the time neighborhood of each node at each moment on the road, and divides each cycle to obtain multiple sub-time periods corresponding to each node in each cycle. The sub-time periods corresponding to different nodes in the same cycle are not necessarily one-to-one corresponding, that is, the traffic conditions at different locations on the road are considered, and then the similarity of traffic characterization values of different nodes at all moments in the same cycle is analyzed. An evaluation was conducted, and the vehicle change evaluation value of each node at each moment was obtained by combining the difference in duration between the sub-time period at each moment and other sub-time periods in its cycle and the discrete degree of the traffic characterization value in the time neighborhood of each node at each moment. Considering that vehicles can reach the location of connected nodes in a relatively short time during driving on the road, the difference in traffic characterization values of each node and its connected nodes at the same moment was analyzed, and the break threshold during clustering was adjusted. The connected graph was clustered based on the adjusted break threshold, which ensured the accuracy of the clustering effect. Furthermore, whether to issue an early warning was determined based on the division results of the clustering clusters, thereby improving the accuracy of the early warning of traffic operation conditions. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0048] Figure 1 The present invention provides a flowchart of a method executed by a ground-to-air integrated operation status monitoring and early warning system according to an embodiment of the present invention. DETAILED DESCRIPTION
[0049] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, a ground-to-air integrated operation status monitoring and early warning system proposed according to the present invention is described in detail below in conjunction with the accompanying drawings and preferred embodiments.
[0050] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
[0051] The specific scheme of the ground-to-air integrated operation status monitoring and early warning system provided by the present invention is described in detail below with reference to the accompanying drawings.
[0052] An embodiment of a ground-to-air integrated operation status monitoring and early warning system:
[0053] The specific scenario targeted by this embodiment is: using a connected graph dynamic split clustering algorithm to process traffic data on the road. Inappropriate setting of the break threshold directly affects the clustering results of subsequent traffic data, which in turn leads to low accuracy in warning of traffic conditions. This embodiment will analyze the collected traffic data at different nodes on the road, adaptively determine the appropriate break threshold, and then obtain accurate clustering results, thereby improving the accuracy of traffic condition warnings.
[0054] This embodiment proposes a ground-to-air integrated operation status monitoring and early warning system. The system is used to achieve Figure 1 The specific steps are as follows:
[0055] Step S1, obtaining traffic data of different nodes on the road at each moment in each cycle in the current time period and location information between different nodes; and constructing a connectivity graph based on the traffic data and location information between different nodes.
[0056] This embodiment first uses GIS satellite technology to obtain traffic data of nodes at different locations on the road in the current time period and location information between different nodes on the road. The connectivity between different nodes can be determined according to the distribution between the nodes. There are many types of traffic information, such as vehicle speed, number of vehicles, etc. This embodiment will take one type of traffic data as an example for explanation. The method provided by this embodiment can be used to process other types of traffic data. This embodiment sets the collection frequency of traffic data to once per second, that is, traffic data of each node on the road is collected every second. The current time period is a collection of all historical moments whose time interval with the current moment is less than or equal to the preset time length. In this embodiment, the preset time length is 10 days. Since the similarity of traffic information in the same time period every day is high, this embodiment sets the duration of each cycle to one day. In specific applications, the implementer can set the preset time length, the collection frequency of traffic data, the number of nodes on the road, and the specific location of the node according to the specific situation. So far, this embodiment has collected traffic data of different nodes on the road at each moment in each cycle in the current time period and the location information of different nodes.
[0057] A connectivity graph is constructed based on the traffic data of different nodes on the road at each moment in each cycle in the current time period and the connectivity between different nodes. It should be noted that the connectivity in this embodiment refers to physical connectivity, and the method for constructing a connectivity graph is an existing technology and will not be elaborated here.
[0058] So far, this embodiment has obtained the traffic data of different nodes on the road at each time in each cycle in the current time period and the corresponding connectivity graph.
[0059] Step S2, obtaining the traffic characterization value of each node at each moment based on the difference between the traffic data of vehicles passing through each node at each moment and the traffic data of the moments in its time neighborhood and the traffic data of vehicles passing through each node at each moment; dividing each cycle based on the traffic characterization value of the moments in the time neighborhood of each moment to obtain the sub-time period corresponding to each node in each cycle.
[0060] When warning the traffic operation status on the road, the biggest problem on the road is generally traffic jam. This embodiment will perform abnormal analysis on the traffic data on the road, and then determine whether to warn the traffic operation status of the road. When performing abnormal analysis on the traffic data on the road, a connectivity graph is constructed for the collected traffic data. The constructed connectivity graph can be divided into clusters by splitting operations. When the traffic data of the collected nodes is constructed, there is certain location information in the traffic network where the nodes themselves are located. Some nodes are physically connected in space. When constructing the connectivity graph, the connectivity graph is constructed by the connectivity of the nodes in the physical sense. The connectivity graph constructed by physical connectivity will be affected by the physical distance. When performing the data splitting operation, the physical distance will affect the division of the data point clustering clusters. At the same time, the physical connectivity will cause vehicles to have certain mobility. When clustering, it is necessary to consider the changes of the acquired traffic data with location and time. By analyzing the changes of the collected traffic data in time and location, the break threshold of the connectivity graph is adjusted to make the acquired abnormal data more accurate. This embodiment will first obtain the traffic characterization value of each node at each moment based on the difference between the traffic data of vehicles passing through each node at each moment and the traffic data of the moments in its time neighborhood and the traffic data of vehicles passing through each node at each moment, and then analyze the traffic characterization value.
[0061] Specifically, for the jth node at the i-th moment:
[0062] The vehicle passing through the jth node at the i-th moment is recorded as the target vehicle at the j-th node at the i-th moment; the traffic characterization value at the j-th node at the i-th moment is calculated using the following formula:
[0063]
[0064] Among them, A ij represents the traffic characterization value of the jth node at the i-th moment, N ij represents the number of target vehicles at the jth node at the i-th moment, N ij represents the number of target vehicles at the jth node at the i-th moment, a ijn represents the traffic data of the nth target vehicle at the i-th time of the j-th node, M represents the number of moments in the time neighborhood of the i-th time, and a' ijn,m represents the traffic data of the nth target vehicle at the jth node at the i-th moment in the time neighborhood of the i-th moment, and exp() represents an exponential function with a natural constant as the base.
[0065] In this embodiment, the total length of the time neighborhood is 10 minutes. For the i-th moment: the i-th moment is taken as the central moment, a time range with a total length of 10 minutes is obtained, and the set of all moments in the time range except the i-th moment is taken as the time neighborhood of the i-th moment. This embodiment obtains the traffic characterization value of the j-th node at the i-th moment by analyzing the difference between the traffic data of the vehicle passing through the j-th node at the i-th moment and the moments in the time neighborhood of the i-th moment.
[0066] By adopting the above method, the traffic characterization value of each node at each moment can be obtained. In order to improve the accuracy of the subsequent clustering results, it is necessary to adjust the clustering in combination with the discrete situation of the traffic characterization value. When analyzing the discrete situation of the obtained traffic characterization value, it is necessary to consider the changing trend of the traffic characterization value with the time series. When analyzing the status of the vehicle, it is necessary to consider the corresponding time situation. Since the traffic data will be affected by the working hours, such as the morning rush hour and the evening rush hour, this embodiment obtains the validity of the corresponding traffic data changing with time by analyzing the consistency of the data changes in each cycle. At the same time, when analyzing the changes of the obtained traffic data with time, it is necessary to consider the influence of the changes of the traffic lights on the road to avoid the situation where the vehicle is driving slowly due to the state of the red light and is mistaken for abnormality.
[0067] When the traffic characterization value changes over time, it is taken into account that the driving conditions of vehicles at different times may be different, that is, the traffic characterization value will have periodic changes, and the traffic data will be affected by the changes in traffic lights. Therefore, this embodiment will divide each cycle into sub-time periods with obvious differences in adjacent data changes based on the traffic characterization values of the moments in the time neighborhood of each node at each moment.
[0068] Specifically, for the i-th moment:
[0069] The moment before the i-th moment in the time neighborhood of the i-th moment is recorded as the first neighboring moment of the i-th moment, and the moment after the i-th moment in the time neighborhood of the i-th moment is recorded as the second neighboring moment of the i-th moment, that is, multiple first neighboring moments and multiple second neighboring moments of the i-th moment are obtained.
[0070] For the j-th node: the sum of the traffic characterization values of all the first neighboring moments of the j-th node at the i-th moment is recorded as the first eigenvalue of the j-th node at the i-th moment; the sum of the traffic characterization values of all the second neighboring moments of the j-th node at the i-th moment is recorded as the second eigenvalue of the j-th node at the i-th moment; the variance of the traffic characterization values of all the first neighboring moments of the j-th node at the i-th moment is recorded as the first variance of the j-th node at the i-th moment; the variance of the traffic characterization values of all the second neighboring moments of the j-th node at the i-th moment is recorded as the second variance of the j-th node at the i-th moment; according to the first eigenvalue, the second eigenvalue, the first square The first eigenvalue and the second eigenvalue are both positively correlated with the segmentation index, and the first variance and the second variance are both negatively correlated with the segmentation index; wherein a positive correlation indicates that the dependent variable will increase as the independent variable increases, and the dependent variable will decrease as the independent variable decreases, and the specific relationship may be a multiplication relationship, an addition relationship, or the idempotence of an exponential function, which is determined by actual application; a negative correlation indicates that the dependent variable will decrease as the independent variable increases, and the dependent variable will increase as the independent variable decreases, which may be a subtraction relationship, a division relationship, etc., which is determined by actual application. As a specific implementation method, a specific calculation formula for the segmentation index of the jth node at the i-th moment is given:
[0071]
[0072] Among them, δ ij represents the segmentation index of the jth node at the i-th moment, ε i,1 represents the variance of the traffic characterization values of all first neighboring nodes at the i-th time of the j-th node, ε i,2 represents the variance of the traffic characterization values of all second neighborhood moments of the j-th node at the i-th moment, M1 represents the number of first neighborhood moments at the i-th moment, M2 represents the number of second neighborhood moments at the i-th moment, norm() represents the normalization function, b1 represents the preset first value, b2 represents the preset second value, and both b1 and b2 are greater than 0.
[0073] In this embodiment, the preset first value and the preset second value are introduced into the calculation formula of the segmentation index to prevent the denominator from being 0. In this embodiment, the preset first value and the preset second value are both 0.1. In specific applications, the implementer can set them according to specific circumstances. represents the first eigenvalue of the jth node at the i-th moment, represents the second eigenvalue of the jth node at the i-th moment, ε i,1 is the first variance, which is used to reflect the discrete degree of the traffic characterization value of all first neighboring moments of the jth node at the i-th moment; εi,2 is the second variance, which is used to reflect the degree of dispersion of the traffic characterization values of all the second neighboring moments of the jth node at the i-th moment. This embodiment evaluates the change of the traffic state of the jth node at the i-th moment by analyzing the difference of the traffic characterization values of the neighboring moments on both sides around the i-th moment and the degree of dispersion of the traffic characterization values of the neighboring moments on each side. The larger it is, the greater the difference between the traffic conditions of the j-th node at the i-th moment and the traffic conditions at other surrounding moments, and the i-th moment should be divided out, that is, the larger the segmentation index of the j-th node at the i-th moment.
[0074] The above method can be used to obtain the segmentation index of each node at each moment.
[0075] For the jth node: the moment when the segmentation index is greater than the preset segmentation threshold is used as the division point, and the moments in each cycle in the current time period are divided based on the division point to obtain the sub-time periods corresponding to the jth node in each cycle. The preset segmentation threshold in this embodiment is 0.7. In specific applications, the implementer can set it according to the specific situation. Using the above method, the moments in each cycle can be divided based on the segmentation index of each node at each moment, and multiple sub-time periods corresponding to each node in each cycle can be obtained. It should be noted that: since the traffic conditions of different nodes at the same time may be different, the segmentation indexes of different nodes at the same time may be different, and the sub-time periods corresponding to different nodes in the same cycle may not correspond one to one, and the number of sub-time periods corresponding to different nodes in the same cycle may be different.
[0076] Step S3, based on the traffic characterization values of each node at all times in each cycle, construct the data sequence corresponding to each node in each cycle, match the data sequences corresponding to different nodes in the same cycle, and obtain the vehicle change evaluation value of each node at each moment according to the matching results, the difference in duration between the sub-time period at each moment and other sub-time periods in the cycle, and the discrete degree of the traffic characterization values of all moments in the time neighborhood of each node at each moment.
[0077] For the jth node: arrange the traffic characterization values of the jth node at all times in each cycle in chronological order, obtain the traffic characterization value sequence of the jth node in each cycle, and record the sequence obtained at this time as the data sequence corresponding to the jth node in each cycle, that is, the jth node corresponds to a data sequence in each cycle. Using this method, the data sequence corresponding to each node in each cycle can be obtained.
[0078] This embodiment has obtained the data sequence corresponding to each node in each cycle, and then uses the dynamic time warping algorithm to match the data sequences corresponding to every two nodes in the same cycle and obtain the matching results. The dynamic time warping algorithm is a prior art and will not be described in detail here.
[0079] For the i-th moment:
[0080] In the cycle where the ith moment is located, with a step length of 2, the ith moment is taken as the starting point, and the corresponding moments are respectively traversed before and after the ith moment to obtain the corresponding moments, and all the corresponding moments obtained are used as the reference moments of the ith moment, that is, multiple reference moments of the ith moment are obtained. The reason why the step length is set to 2 in this embodiment is that the influence of traffic lights on the traffic data of nodes on the road is taken into consideration.
[0081] For the jth node: if the number of matching points of the traffic characterization value of the jth node at the i-th moment in the matching results of the data sequence corresponding to the jth node in the cycle at the i-th moment and the data sequence corresponding to the r-th node in the cycle at the i-th moment is 1, that is, a one-to-one situation, then the matching point is used as the matching point corresponding to the jth node in the r-node; if the number of matching points of the traffic characterization value of the jth node at the i-th moment in the matching results of the data sequence corresponding to the j-th node in the cycle at the i-th moment and the data sequence corresponding to the r-th node in the cycle at the i-th moment is greater than 1, that is, a one-to-many situation, then the matching point with the shortest time interval with the i-th moment among all the matching points is used as the matching point corresponding to the j-th node in the r-node; using the above method, the matching point corresponding to the j-th node in each node can be obtained. Next, the vehicle change evaluation value of the jth node at the i-th moment is obtained based on the difference in duration between the sub-time period at the i-th moment and the sub-time period at the q-th reference moment at the i-th moment, the traffic characterization value of the j-th node at the i-th moment, the traffic characterization value of the matching point corresponding to the j-th node in the r-th node, and the variance of the traffic characterization values of all moments in the neighborhood of the j-th node at the i-th moment. The specific calculation formula for the vehicle change evaluation value of the j-th node at the i-th moment is:
[0082]
[0083] Among them, W ij represents the vehicle change evaluation value of the jth node at the i-th moment, Q represents the number of reference moments at the i-th moment, and l i Indicates the duration of the sub-time period at the i-th moment, l i,q represents the duration of the sub-time period where the qth reference moment of the i-th moment is located, R represents the number of nodes, and f ijrepresents the traffic characterization value of the j-th node at the i-th moment, represents the traffic characterization value of the matching point corresponding to the jth node in the rth node, ε i,j represents the variance of the traffic characterization values at all times in the neighborhood of the j-th node at the i-th time, || represents the absolute value sign, and exp() represents an exponential function with a natural constant as the base.
[0084] The variance of the traffic characterization values at all times in the neighborhood of the jth node at the i-th time is used to reflect the degree of dispersion of the traffic characterization values at all times in the neighborhood of the jth node at the i-th time. The larger the variance, the greater the degree of dispersion of the traffic characterization values at all times in the neighborhood of the jth node at the i-th time. i -|l i,q | is used to reflect the difference in duration between the sub-time period of the ith moment and the sub-time period of the qth reference moment of the ith moment. The smaller the difference in duration, the more likely the change in traffic conditions within the sub-time period is to change with the change of traffic lights; It is used to reflect the difference between the traffic characterization value of the jth node at the i-th moment and the traffic characterization value of the matching point corresponding to the jth node in the rth node. The larger the difference, the more special the traffic condition of the jth node at the i-th moment, and the more invalid the traffic characterization value of the jth node at the i-th moment is when performing trend analysis. The greater the discreteness of the traffic characterization values of all moments in the neighborhood of the jth node at the i-th moment, the greater the difference in the length of the sub-time period where the i-th moment is located and the sub-time period where the reference moment of the i-th moment is located, and the greater the difference between the traffic characterization value of the jth node at the i-th moment and the traffic characterization value of the matching point corresponding to the j-th node in each node, the greater the vehicle change evaluation value of the jth node at the i-th moment.
[0085] By adopting the above method, the vehicle change evaluation value at each node at each moment can be obtained.
[0086] Step S4, according to the difference in traffic characterization values of each node and the nodes connected to it at the same time and the vehicle change evaluation value, the fracture threshold corresponding to each node is adjusted to obtain the adjusted fracture threshold corresponding to each node at each time, the connected graph is divided into clusters based on the adjusted fracture threshold, and whether to warn is determined based on the cluster division result.
[0087] In step S3, the present embodiment has obtained the vehicle change evaluation value of each node at each moment. The position information of the vehicle will change with time during the movement. According to the physical connectivity between different nodes in the obtained connectivity graph, the fracture threshold corresponding to each node is adjusted to make the subsequent clustering results more accurate. Therefore, according to the difference in traffic characterization value of each node at the same moment with the node connected to it and the vehicle change evaluation value of the node connected to each node at each moment, the present embodiment will adjust the fracture threshold corresponding to each node, and then obtain the adjusted fracture threshold corresponding to each node at each moment.
[0088] Specifically, for the jth node: arrange the vehicle change evaluation values of the jth node at all moments in each cycle in chronological order, and obtain the vehicle change evaluation value sequence corresponding to the jth node in each cycle, that is, the jth node corresponds to a vehicle change evaluation value sequence in each cycle. By adopting this method, the vehicle change evaluation value sequence corresponding to each node in each cycle can be obtained, and the vehicle change evaluation value sequences corresponding to each two nodes in the same cycle are matched by the dynamic time warping algorithm, and the corresponding matching data are obtained; it should be noted that: when matching two vehicle change evaluation value sequences, there will be a one-to-many matching situation, that is, a certain data corresponds to multiple matching data. If this situation occurs, this embodiment will use the vehicle change evaluation value at the moment with the shortest time interval as the matching data corresponding to the data. The dynamic time warping algorithm is a prior art and will not be described in detail here. For the jth node at the i-th moment: the node connected to the jth node is recorded as the reference node of the jth node; the connection in this embodiment refers to physical connection, that is, if the vehicle can directly reach another node through one node without passing through any other nodes in between, then one of the two nodes is determined to be a node connected to the other node. Therefore, for the jth node, there may be multiple reference nodes. For the h-th reference node: the difference between the traffic characterization value of the j-th node at the i-th moment and the traffic characterization value of the matching point corresponding to the j-th node in the h-th node is recorded as the first difference corresponding to the h-th reference node of the j-th node; the specific method for obtaining the difference between the traffic characterization value of the j-th node at the i-th moment and the traffic characterization value of the matching point corresponding to the j-th node in the h-th node is: the absolute value of the difference between the traffic characterization value of the j-th node at the i-th moment and the traffic characterization value of the matching point corresponding to the j-th node in the h-th node is recorded as the difference between the traffic characterization value of the j-th node at the i-th moment and the traffic characterization value of the matching point corresponding to the k-th reference node of the j-th node in the r-th node. By adopting the above method, the first difference corresponding to each reference node of the j-th node can be obtained, and each reference node of the j-th node corresponds to a first difference.The sum of the first differences corresponding to all reference nodes of the jth node is recorded as the third eigenvalue of the jth node; the variance of the matching data of the vehicle change evaluation values of all reference nodes of the jth node at the i-th moment is calculated and recorded as the third variance; the adjustment parameter of the jth node at the i-th moment is obtained according to the third eigenvalue and the third variance, the third variance is positively correlated with the adjustment parameter, and the third eigenvalue is negatively correlated with the adjustment parameter; wherein the positive correlation indicates that the dependent variable will increase with the increase of the independent variable, and the dependent variable will decrease with the decrease of the independent variable, and the specific relationship can be a multiplication relationship, an addition relationship, and the idempotence of an exponential function, which is determined by actual application; the negative correlation indicates that the dependent variable will decrease with the increase of the independent variable, and the dependent variable will increase with the decrease of the independent variable, which can be a subtraction relationship, a division relationship, etc., which is determined by actual application. As a specific implementation method, a specific calculation formula for the adjustment parameter of the jth node at the i-th moment is given:.
[0089]
[0090] in, represents the adjustment parameter of the jth node at the i-th moment, ε' i,j represents the variance of the matching data of the vehicle change evaluation values of all reference nodes of the jth node at the i-th moment, H represents the number of reference nodes of the jth node, Δf ijh represents the first difference corresponding to the rth reference node of the jth node, norm() represents the normalization function, λ is the preset anti-zero constant, and λ is greater than 0.
[0091] The purpose of introducing the preset anti-zero constant into the calculation formula of the adjustment parameter is to prevent the denominator from being 0. In this embodiment, the value of the preset anti-zero constant is 1. In specific applications, the implementer can set it according to the specific situation. i,j is the third variance, which is used to reflect the degree of discreteness of the matching data of the vehicle change evaluation values of all reference nodes of the j-th node at the i-th moment. The larger the variance, the greater the degree of discreteness of the matching data of the vehicle change evaluation values of all reference nodes of the j-th node at the i-th moment, which means that the j-th node is more likely to be affected by multiple connected nodes at the i-th moment. When clustering, it is more likely to appear as discrete data, the j-th node is more likely to be a traffic hub, and its break threshold should be increased; the larger the first difference corresponding to each reference node of the j-th node, the more inconsistent the change of the acquired data is, the more likely the traffic data anomaly of the j-th node at the i-th moment is caused by an unexpected situation, which does not belong to the overall change, and the break threshold should be reduced when clustering.
[0092] The sum of the adjustment parameter of the j-th node at the i-th moment and the constant 1 is recorded as the adjustment coefficient of the j-th node at the i-th moment, and the product of the adjustment coefficient and the initial fracture threshold is used as the adjusted fracture threshold corresponding to the j-th node at the i-th moment. The specific calculation formula of the adjusted fracture threshold corresponding to the j-th node at the i-th moment is: Among them, Y i represents the adjusted fracture threshold corresponding to the j-th node at the i-th moment, and y represents the initial fracture threshold. In specific applications, the implementer sets the initial fracture threshold according to the specific situation.
[0093] By adopting the above method, the adjusted break threshold corresponding to each node at each time can be obtained. Based on the adjusted break threshold, the connected graph dynamic split clustering algorithm is used to divide the clusters in the connected graph to obtain multiple clusters. The connected graph dynamic split clustering algorithm is a prior art and will not be described in detail here.
[0094] For any cluster: if the number of data points in the cluster is less than the preset first number, the cluster is regarded as an abnormal cluster; the above method can be used to screen out abnormal clusters from all clusters. In this embodiment, the preset first number is 50. In specific applications, the implementer can set it according to the specific situation. Calculate the sum of the number of data points in all abnormal clusters and the number of all data points that are not in the cluster; if the sum is greater than the preset second number, an early warning is issued to remind the traffic police to intervene in the driving conditions of vehicles on the road in time; if the sum is less than or equal to the preset second number, no early warning is issued; the preset first number is less than the preset second number. In this embodiment, the preset second number is 300. In specific applications, the implementer can set it according to the specific situation.
[0095] Thus, by adopting the method provided in this embodiment, when the traffic operation condition is abnormal, early warning can be given in time, and the abnormal traffic operation can be regulated by manual intervention.
[0096] This embodiment takes into account that when the connected graph dynamic splitting clustering algorithm is used to cluster the traffic data on the road, inappropriate setting of the fracture threshold will directly affect the clustering effect, thereby resulting in low accuracy of warning of traffic operation conditions and high similarity of traffic conditions on the same road section in the same time period every day. Therefore, this embodiment divides each cycle based on the traffic characterization value of each moment in the time neighborhood of each node at each moment on the road to obtain multiple sub-time periods corresponding to each node in each cycle. The sub-time periods corresponding to different nodes in the same cycle are not necessarily one-to-one corresponding, that is, the traffic conditions at different locations on the road are considered, and then the similarity of the traffic characterization values of different nodes at all moments in the same cycle is calculated. An evaluation was conducted, and the vehicle change evaluation value of each node at each moment was obtained by combining the difference in duration between the sub-time period at each moment and other sub-time periods in its cycle and the discrete degree of the traffic characterization value in the time neighborhood of each node at each moment. Considering that vehicles can reach the locations of connected nodes in a shorter time during driving on the road, the differences in traffic characterization values of each node and its connected nodes at the same moment were analyzed, and the break threshold during clustering was adjusted. The connectivity graph was clustered based on the adjusted break threshold to ensure the accuracy of the clustering effect. Furthermore, whether to issue an early warning is determined based on the division results of the clustering clusters, thereby improving the accuracy of the early warning of traffic operation conditions.
[0097] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A ground-to-air integrated operation status monitoring and early warning system, comprising a memory and a processor, characterized in that: The processor executes the computer program stored in the memory to implement the following steps: Obtaining traffic data of different nodes on the road at each moment in each cycle in the current time period and location information between different nodes; constructing a connectivity graph based on the traffic data and location information between different nodes; According to the difference between the traffic data of the vehicles passing through each node at each moment and the traffic data of the vehicles passing through each node at each moment, the traffic characterization value of each node at each moment is obtained; based on the traffic characterization value of the moment in the time neighborhood of each node at each moment, each cycle is divided to obtain the sub-time period corresponding to each node in each cycle; Based on the traffic characterization value of each node at all times in each cycle, the data sequence corresponding to each node in each cycle is constructed, and the data sequences corresponding to different nodes in the same cycle are matched. According to the matching results, the difference in the length of the sub-time period at each moment and other sub-time periods in the cycle, and the discrete degree of the traffic characterization value of all moments in the time neighborhood of each node at each moment, the vehicle change evaluation value of each node at each moment is obtained; According to the difference in traffic characterization values of each node and the nodes connected to it at the same time and the vehicle change evaluation value, the fracture threshold corresponding to each node is adjusted to obtain the adjusted fracture threshold corresponding to each node at each time, the connected graph is divided into clusters based on the adjusted fracture threshold, and whether to issue an early warning is determined based on the cluster division result.
2. The ground-to-air integrated operation status monitoring and early warning system according to claim 1 is characterized in that: The method of obtaining a traffic characterization value of each node at each moment according to the difference between the traffic data of the vehicles passing through each node at each moment and the traffic data of the vehicles passing through each node at each moment in its time neighborhood includes: For the jth node at the i-th moment: The vehicle passing through the jth node at the i-th moment is recorded as the target vehicle at the j-th node at the i-th moment; The traffic characterization value of the jth node at the i-th moment is calculated using the following formula: Among them, A ij represents the traffic characterization value of the jth node at the i-th moment, N ij represents the number of target vehicles at the jth node at the i-th moment, a ijn represents the traffic data of the nth target vehicle at the i-th time of the j-th node, M represents the number of time moments in the time neighborhood of the i-th time moment, and a′ ijn,m represents the traffic data of the nth target vehicle at the jth node at the i-th moment in the time neighborhood of the i-th moment, and exp() represents an exponential function with a natural constant as the base.
3. The ground-to-air integrated operation status monitoring and early warning system according to claim 1 is characterized in that: The method of dividing each cycle based on the traffic characterization value of each node at each time in the time neighborhood to obtain the sub-time period corresponding to each node in each cycle includes: For the i-th moment: The moment before the i-th moment in the time neighborhood of the i-th moment is recorded as the first neighboring moment of the i-th moment, and the moment after the i-th moment in the time neighborhood of the i-th moment is recorded as the second neighboring moment of the i-th moment; For the jth node: the sum of the traffic characterization values of all the first neighboring moments of the jth node at the i-th moment is recorded as the first eigenvalue of the jth node at the i-th moment; the sum of the traffic characterization values of all the second neighboring moments of the jth node at the i-th moment is recorded as the second eigenvalue of the jth node at the i-th moment; the variance of the traffic characterization values of all the first neighboring moments of the jth node at the i-th moment is recorded as the first variance of the jth node at the i-th moment; the variance of the traffic characterization values of all the second neighboring moments of the jth node at the i-th moment is recorded as the second variance of the jth node at the i-th moment; according to the first eigenvalue, the second eigenvalue, the first variance and the second variance, the segmentation index of the jth node at the i-th moment is obtained, the first eigenvalue and the second eigenvalue are both positively correlated with the segmentation index, and the first variance and the second variance are both negatively correlated with the segmentation index; Based on the segmentation index of each node at each moment, the moments in the cycle within the current time period are divided to obtain the sub-time periods corresponding to each node in each cycle.
4. The ground-to-air integrated operation status monitoring and early warning system according to claim 3 is characterized in that: The step of dividing the moments in the cycle in the current time period based on the segmentation index of each node at each moment to obtain the sub-time periods corresponding to each node in each cycle includes: For the j-th node: the moment when the segmentation index is greater than the preset segmentation threshold is taken as the division point, and the moments in each cycle in the current time period are divided based on the division point to obtain the sub-time periods corresponding to the j-th node in each cycle.
5. The ground-to-air integrated operation status monitoring and early warning system according to claim 1 is characterized in that: The dynamic time warping algorithm is used to match the data sequences corresponding to different nodes in the same period.
6. The ground-to-air integrated operation status monitoring and early warning system according to claim 5 is characterized in that: The vehicle change evaluation value of each node at each moment is obtained according to the matching results, the difference in the length of the sub-time period at each moment and other sub-time periods in its cycle, and the discrete degree of the traffic characterization values of all moments in the time neighborhood of each node at each moment, including: For the i-th moment: In the cycle where the i-th moment is located, with a step length of 2, take the i-th moment as the starting point, traverse before and after the i-th moment respectively, obtain the corresponding moments, and use all the corresponding moments obtained as the reference moments of the i-th moment; For the jth node: if the number of matching points of the traffic characterization value of the jth node at the i-th moment in the matching results of the data sequence corresponding to the jth node in the cycle at the i-th moment and the data sequence corresponding to the rth node in the cycle at the i-th moment is 1, then the matching point is used as the matching point corresponding to the jth node in the r-th node; if the number of matching points of the traffic characterization value of the jth node at the i-th moment in the matching results of the data sequence corresponding to the jth node in the cycle at the i-th moment and the data sequence corresponding to the r-th node in the cycle at the i-th moment is greater than 1, then the matching point with the shortest time interval with the i-th moment among all the matching points is used as the matching point corresponding to the jth node in the r-th node; According to the difference in duration between the sub-time period at the i-th moment and the sub-time period at the q-th reference moment of the i-th moment, the traffic characterization value of the j-th node at the i-th moment, the traffic characterization value of the matching point corresponding to the j-th node in the r-th node, and the variance of the traffic characterization values of all moments in the neighborhood of the j-th node at the i-th moment, the vehicle change evaluation value of the j-th node at the i-th moment is obtained.
7. The ground-to-air integrated operation status monitoring and early warning system according to claim 6 is characterized in that: The following formula is used to calculate the vehicle change evaluation value at the jth node at the i-th moment: Among them, W ij represents the vehicle change evaluation value of the jth node at the i-th moment, Q represents the number of reference moments at the i-th moment, and l i Indicates the duration of the sub-time period at the i-th moment, l i,q represents the duration of the sub-time period where the qth reference moment of the i-th moment is located, R represents the number of nodes, and f ij represents the traffic characterization value of the j-th node at the i-th moment, represents the traffic characterization value of the matching point corresponding to the jth node in the rth node, ε i,j represents the variance of the traffic characterization values at all times in the neighborhood of the j-th node at the i-th time, || represents the absolute value sign, and exp() represents an exponential function with a natural constant as the base.
8. The ground-to-air integrated operation status monitoring and early warning system according to claim 6 is characterized in that: The method of adjusting the fracture threshold corresponding to each node according to the difference in traffic characterization value between each node and the node connected to it at the same time and the vehicle change evaluation value to obtain the adjusted fracture threshold corresponding to each node at each time includes: Based on the vehicle change evaluation values of each node at all times in each cycle, a vehicle change evaluation value sequence corresponding to each node in each cycle is constructed; a dynamic time warping algorithm is used to match the vehicle change evaluation value sequences corresponding to different nodes in the same cycle to obtain matching data of the vehicle change evaluation value of each node at each time; For the jth node at the i-th moment: The node connected to the jth node is recorded as the reference node of the jth node; For the h-th reference node: the difference between the traffic characterization value of the j-th node at the i-th moment and the traffic characterization value of the matching point corresponding to the j-th node in the h-th node is recorded as the first difference corresponding to the h-th reference node of the j-th node; The sum of the first differences corresponding to all reference nodes of the j-th node is recorded as the third eigenvalue of the j-th node; the variance of the matching data of the vehicle change evaluation values of all reference nodes of the j-th node at the i-th moment is calculated, recorded as the third difference; Obtaining an adjustment parameter of the j-th node at the i-th moment according to the third eigenvalue and the third difference, wherein the third difference is positively correlated with the adjustment parameter, and the third eigenvalue is negatively correlated with the adjustment parameter; An adjusted fracture threshold corresponding to the j-th node at the i-th moment is determined based on the adjustment parameter.
9. The ground-to-air integrated operation status monitoring and early warning system according to claim 8 is characterized in that: The determining, based on the adjustment parameter, the adjusted fracture threshold corresponding to the j-th node at the i-th moment includes: The sum of the adjustment parameter of the jth node at the i-th moment and the constant 1 is recorded as the adjustment coefficient of the jth node at the i-th moment, and the product of the adjustment coefficient and the initial fracture threshold is taken as the adjusted fracture threshold corresponding to the jth node at the i-th moment.
10. The ground-to-air integrated operation status monitoring and early warning system according to claim 1 is characterized in that: The step of judging whether to issue an early warning based on the clustering result includes: For any cluster: if the number of data points in the cluster is less than a preset first number, the cluster is regarded as an abnormal cluster; Calculate the sum of the number of data points in all abnormal clusters and the number of data points not in the clusters; if the sum is greater than a preset second number, issue a warning; if the sum is less than or equal to the preset second number, do not issue a warning; the preset first number is less than the preset second number.
